This article really doesn't have that much to do with the Turing test for most of its extent. The point of the Turing test isn't merely that under some circumstances machines can be confused with humans. The whole point of the Turing test is that it takes something that we think is essential to being intelligent or being conscious, and has the machine replicating that exactly. Or at least, that's how Turing intended it. Building sophisticated mannequins doesn't cut it - hopefully no-one thinks that merely looking like a human being means that something is intelligent and/or conscious, no matter how good they look.
(if the automata Ishiguro has produced also answer students' questions intelligently, then the situation is different, and more like that of the original Turing test).
Similarly, what opponents humans like to play against doesn't really show us anything. If anything, it shows that Turing-like tests are unreliable because people tend to think that if it has a human face it's more 'interesting'.
The business with computers and war crimes simply begs the question. Certainly a computer that releases nuclear missiles in response to certain conditions is possible now - in fact, I'd be surprised if such a dead-man's switch doesn't exist already. But even though no person has pulled the switch doesn't mean that the computer is guilty of war crimes, any more than a car left in gear that runs over someone is guilty of a traffic violation. You need rationality and knowledge of consequences of actions to be guilty of anything. Note also that there's nothing about deceiving humans here, so I'm not really sure why this is even in the same article.
The only thing in here which comes close to something that sounds anything like intelligence (let alone consciousness) is the jazz-playing robot.
So it's a bit of a hodgepodge mess of completely different issues. It would be better (but less breathlessly exciting) to take out the stuff about war-crimes and any mention of the Turing test and call it 'computers with human faces' or something. At least then it would have a unified subject-matter.
copying is one thing --- saying you wrote something that someone else did is completely different. One's copyright infringement (or 'theft'), the other is fraud.
(I suspect that this might be the way it would go legally, also, but of course IAN... you get the idea)
there's plenty of people who don't like copyright, especially in its current form, but although some of them might well support copying everything and giving it to your friends as a good thing to do, I don't think a single one would be supportive of deliberately misattributing someone else's work.
A) When you reduce the the N dimensional space, your would start by eliminating noise words (ie words that only occur in a single email). Spam emails that put fake words in to lower their spam probability in the bayesian method would not benefit with this method.
The method described by Paul Graham only looks at a handful of the most 'interesting' words in the mail ('interesting' meaning tending to yield a high probability of being either spam or ham). Adding lots of random words could mean that the spammer lucks out and gets words that happen to appear a lot in your regular mail, but what's rather more likely is that the 'interesting' words will be things like 'viagra', and the random words
will simply be ignored. Bayesian sorting isn't necessarily particularly vunerable to random words.
What would tend to defeat Graham's filter more would be inconsistent spelling of key words, i.e. v1agra, v|agra, V!agr4 or whatever. Perhaps other bayesian filters are cleverer.
This is a strange argument to a strange conclusion.
Whether or not Debian's package management system is better has little enough to do with how fanatics behave. They will tell you it's far superior, no matter what the truth of the matter is.
And how can becoming technically inferior relative to the other distributions possibly be good on balance at all, let alone the 'biggest good'?
You can replace 'Debian' with 'Linux', and 'other distros' with 'Windows' throughout, and you're left with the same argument, and the conclusion being that the best thing Linux could do is to focus on not having any advantages over Windows, as that would reduce the ammount of adolescent zealotous behaviour.
But if so, I think it does rather validate the 'viral license' accusation, although not quite in the way MS seemed to be intending it to be taken.
Here's the scenario: your company, a medium-sized software business, has a few commercial projects underway. One of your employees naughitly includes some of the code from the commercial project into some GPL'd project, which later makes its way into RedHat. Let's say over a few years she steals quite some bucketful of code. You don't notice, of course, because checking all your code against all of RedHat's source would be kinda onerous and not something that would occur to most people.
Then your company starts distributing RedHat. Then you discover that you've been distributing your own code, inadvertantly, under the GPL, so there's nothing that you can do except fire and sue your employee.
'Tough Cheese', you may respond. But if this is right, then it's going to make many companies quite wary about having anything to do with the GPL.
I guess you could always just tip the FSF off quietly. If it's as obvious as you say, all it would take would be to point them towards the download site or whatever.
I suppose if you're already making noises about GPL compliance, you'll be the first under suspicion if the FSF indicates that they've been tipped off. But it's a bit safer than saying "I'm going to tell the FSF *right now*', and you could possibly threaten action if they fire you on a mere suspicion.
I seem to recall there was a cheap Terry Pratchett nock-off (same cover illustrator) who had the Seven Other Dwarves whose names were things like "Sleazy, Stupid, 'Fraidy" and so on.
Well, that might be the only important remedy as far as making the world a nicer place to write software is concerned.
There is also the issue of punishment, of course. If microsoft (or anyone else) has indeed involved itself in evil (and illegal) acts, then it's generally accepted that there is this issue, both as some kind of retribution and as a disincentive for them or anyone else to behave similarly in the future.
While I wouldn't go as far as 'very poor' (I've read too much crap to use such strong language for something that's only mildly disapointing) I agree.
The 'future credibility' paragraph is a good example. VinodV is admittedly somewhat vauge in that paragraph, but he could well be talking about a *major* feature addition, not just the addition of a new command line option to display errors in capitals.
The LWN author doesn't really address this issue at all, just mumbles something about developers' statments being either careful or unreliable, and blithers about MS code being hidden from public view, neither of which seem particularly relevent to the issue of whether OSS *can* produce genuinely new features, or just improve the ones already in the bag.
Of course, there *are* examples of genuine new features comming out of Free Software, but it's also true that it spends much of its time reproducing features available elsewhere. As far as I can tell, most of the end-user applications, including the GNU utilities, are improved versions of older software (in some cases much improved).
OTOH, it's not like commercial software is known for it's revolutionaryness in general, either.
Are there any examples of a genuinely new feature in a Free Software project intended on the end-user?
anyone who cares about swear words should be in favour of their censorship. They're the only words in Western Society with a sacred or magical quality (if you don't believe me, swear at your mother and see how a couple of monosyllabic words can entirely change her demeanor).
Of course, if people ever come to think that these are just words then their magic will be lost. And this is exactly what will happen if they become commonplace in domain names.
I'm sorry for the people who can't make jokes with domain names, but sacrifices must be made by all to preserve the unique culture of swearing.
support censorship of swearing today, to protect the might of swearing for tommorrow.
or something, as the whole work is stylistically very reminiscent of Wittgenstien's Tractatus. The numbered propositions, the use of analogy, the parenthetical comments, it's all there.
Compare:
44. The point of lifestreams isn't to shift from one software structure to another but to shift the whole premis to computerized information: to stop building glorified file cabinets and start building (simplified, abstract) artificial minds; and to store our electronic lives inside.
4.014 A gramaphone record, the musical scale, the written notes, and the sound waves, all stand to one antoher in the same internal relation of depicting that holds between language and the world. they are all constructed according to a common logical pattern. (like the two youths in the fairy-tale, their two horses, and their lilies. They are all in a certain sense one)
I'm not uninterested in Amiga. I'd like to know more about why it was so good and whether/why it will be again (I missed out the first time around). But in this case, the slashdot comments have generally been much more contentful than the orginal article, which was one man's not-too-coherent testiment of faith (if you blacked out the word Amiga could you really tell what he was talking about?)
which is, of course, the opposite of what generally happens on Slashdot. Kinda sad, really.
A stereotypical Amigan is fanatical, loudmouthed, doesn't shut up about the Amiga, somewhat like me, I'm a perfect example. If you can possibly realize why we are so like that, you can see. It's like having this brilliantly cool toy, and nobody else can experience it like you have. It's almost like having a divine experience to search for words, something like this. You've got to tell everybody about it. They may not understand, but if you can get through to one more person, wow. That's them helped out, especially in the world of the Microsoft monopoly.
This statement and others like it from the article are no different from the experience of any enthusiasts dedicated to something outside the mainstream, whether it be Amiga, Linux, Karl Popper or Tottenham Hotspur.
The phenomenom is interesting in its own right, of course, but --- as seen from the outside --- the article gives no compelling reasons to be interested in the content of this particular instance of fanaticism.
I'm afraid I'm not overly impressed with the reports of the conference.
The topics seem to confuse machine consciousness with dangers of (self-replicating) nanotechnology. These are quite seperate topics in my opinion, as seperate as "is heavier than air flight possible" and "will internal combustion engines destroy the environment". Internal combustion engines are neither necessary nor sufficient for flight, and at the very speculative nature of the discussion at this stage concern with secondary effects of putting a particular implementation into practive only serves to muddy the waters.
Basically, the general problem is one of complete ignorance of the terms of the debate. The human brain is the only object in the universe which is uncontentiously involved with consciousness, and we haven't the slightest idea how it works.
Our ignorance of what consciousness is exactly is even more profound, and there are good reasons to suppose that consciousness is one of the great unanswerables, as it is the branch upon which we sit to say anything at all. To turn the saw of your analysis onto the branch that you're sitting on might only result in a fall.
There is no gaurantee at this stage that consciousness can be supported by a universal turing machine (and thus not by a computer). Hope in faster processors might be as folorn as hope in faster steam engines. Even if it could, all attempts so far at producing a viable program have been rather disappointing.
the AI community have been promising machine consciousness "in the next decade or two" for about thirty years now, and it doesn't look like we've got very far.
"IT WAS A PLEASURE TO DO BUSINESS WITH ONE OF THE BIGGEST E-COMMERCE GIANTS IN THE WORLD: F*U*C*K*I*N*G* INTENSE & REALLY EXPENSIVE BUT A KICK FOR ALL OF US! WE ALL LEARNED A LOT"
Slashdot Mirror
This article really doesn't have that much to do with the Turing test for most of its extent. The point of the Turing test isn't merely that under some circumstances machines can be confused with humans. The whole point of the Turing test is that it takes something that we think is essential to being intelligent or being conscious, and has the machine replicating that exactly. Or at least, that's how Turing intended it. Building sophisticated mannequins doesn't cut it - hopefully no-one thinks that merely looking like a human being means that something is intelligent and/or conscious, no matter how good they look.
(if the automata Ishiguro has produced also answer students' questions intelligently, then the situation is different, and more like that of the original Turing test).
Similarly, what opponents humans like to play against doesn't really show us anything. If anything, it shows that Turing-like tests are unreliable because people tend to think that if it has a human face it's more 'interesting'.
The business with computers and war crimes simply begs the question. Certainly a computer that releases nuclear missiles in response to certain conditions is possible now - in fact, I'd be surprised if such a dead-man's switch doesn't exist already. But even though no person has pulled the switch doesn't mean that the computer is guilty of war crimes, any more than a car left in gear that runs over someone is guilty of a traffic violation. You need rationality and knowledge of consequences of actions to be guilty of anything. Note also that there's nothing about deceiving humans here, so I'm not really sure why this is even in the same article.
The only thing in here which comes close to something that sounds anything like intelligence (let alone consciousness) is the jazz-playing robot.
So it's a bit of a hodgepodge mess of completely different issues. It would be better (but less breathlessly exciting) to take out the stuff about war-crimes and any mention of the Turing test and call it 'computers with human faces' or something. At least then it would have a unified subject-matter.
copying is one thing --- saying you wrote something that someone else did is completely different. One's copyright infringement (or 'theft'), the other is fraud.
(I suspect that this might be the way it would go legally, also, but of course IAN... you get the idea)
there's plenty of people who don't like copyright, especially in its current form, but although some of them might well support copying everything and giving it to your friends as a good thing to do, I don't think a single one would be supportive of deliberately misattributing someone else's work.
What would tend to defeat Graham's filter more would be inconsistent spelling of key words, i.e. v1agra, v|agra, V!agr4 or whatever. Perhaps other bayesian filters are cleverer.
It's an obvious name for an X successor, of course, but there is (well, was) a Y Windowing System in development by the Hungry Progammers
Nah, it's all the Dungeons and Dragons weenies in Microsoft.
This is a strange argument to a strange conclusion.
Whether or not Debian's package management system is better has little enough to do with how fanatics behave. They will tell you it's far superior, no matter what the truth of the matter is.
And how can becoming technically inferior relative to the other distributions possibly be good on balance at all, let alone the 'biggest good'?
You can replace 'Debian' with 'Linux', and 'other distros' with 'Windows' throughout, and you're left with the same argument, and the conclusion being that the best thing Linux could do is to focus on not having any advantages over Windows, as that would reduce the ammount of adolescent zealotous behaviour.
That may well be the point of view of the law.
But if so, I think it does rather validate the 'viral license' accusation, although not quite in the way MS seemed to be intending it to be taken.
Here's the scenario: your company, a medium-sized software business, has a few commercial projects underway. One of your employees naughitly includes some of the code from the commercial project into some GPL'd project, which later makes its way into RedHat. Let's say over a few years she steals quite some bucketful of code. You don't notice, of course, because checking all your code against all of RedHat's source would be kinda onerous and not something that would occur to most people.
Then your company starts distributing RedHat. Then you discover that you've been distributing your own code, inadvertantly, under the GPL, so there's nothing that you can do except fire and sue your employee.
'Tough Cheese', you may respond. But if this is right, then it's going to make many companies quite wary about having anything to do with the GPL.
Which I think would be bad.
I guess you could always just tip the FSF off quietly. If it's as obvious as you say, all it would take would be to point them towards the download site or whatever.
I suppose if you're already making noises about GPL compliance, you'll be the first under suspicion if the FSF indicates that they've been tipped off. But it's a bit safer than saying "I'm going to tell the FSF *right now*', and you could possibly threaten action if they fire you on a mere suspicion.
A.
I seem to recall there was a cheap Terry Pratchett nock-off (same cover illustrator) who had the Seven Other Dwarves whose names were things like "Sleazy, Stupid, 'Fraidy" and so on.
Well, that might be the only important remedy as far as making the world a nicer place to write software is concerned.
There is also the issue of punishment, of course. If microsoft (or anyone else) has indeed involved itself in evil (and illegal) acts, then it's generally accepted that there is this issue, both as some kind of retribution and as a disincentive for them or anyone else to behave similarly in the future.
While I wouldn't go as far as 'very poor' (I've read too much crap to use such strong language for something that's only mildly disapointing) I agree.
The 'future credibility' paragraph is a good example. VinodV is admittedly somewhat vauge in that paragraph, but he could well be talking about a *major* feature addition, not just the addition of a new command line option to display errors in capitals.
The LWN author doesn't really address this issue at all, just mumbles something about developers' statments being either careful or unreliable, and blithers about MS code being hidden from public view, neither of which seem particularly relevent to the issue of whether OSS *can* produce genuinely new features, or just improve the ones already in the bag.
Of course, there *are* examples of genuine new features comming out of Free Software, but it's also true that it spends much of its time reproducing features available elsewhere. As far as I can tell, most of the end-user applications, including the GNU utilities, are improved versions of older software (in some cases much improved).
OTOH, it's not like commercial software is known for it's revolutionaryness in general, either.
Are there any examples of a genuinely new feature in a Free Software project intended on the end-user?
Of course, if people ever come to think that these are just words then their magic will be lost. And this is exactly what will happen if they become commonplace in domain names.
I'm sorry for the people who can't make jokes with domain names, but sacrifices must be made by all to preserve the unique culture of swearing.
support censorship of swearing today, to protect the might of swearing for tommorrow.
Compare:
44. The point of lifestreams isn't to shift from one software structure to another but to shift the whole premis to computerized information: to stop building glorified file cabinets and start building (simplified, abstract) artificial minds; and to store our electronic lives inside.
4.014 A gramaphone record, the musical scale, the written notes, and the sound waves, all stand to one antoher in the same internal relation of depicting that holds between language and the world.
they are all constructed according to a common logical pattern.
(like the two youths in the fairy-tale, their two horses, and their lilies. They are all in a certain sense one)
(apologies in advance for the poor latin).
>For then, why did you read it?
I'm not uninterested in Amiga. I'd like to know more about why it was so good and whether/why it will be again (I missed out the first time around). But in this case, the slashdot comments have generally been much more contentful than the orginal article, which was one man's not-too-coherent testiment of faith (if you blacked out the word Amiga could you really tell what he was talking about?)
which is, of course, the opposite of what generally happens on Slashdot. Kinda sad, really.
(thanks for the URL, BTW)
A stereotypical Amigan is fanatical, loudmouthed, doesn't shut up about the Amiga, somewhat like me, I'm a perfect example. If you can possibly realize why we are so like that, you can see. It's like having this brilliantly cool toy, and nobody else can experience it like you have. It's almost like having a divine experience to search for words, something like this. You've got to tell everybody about it. They may not understand, but if you can get through to one more person, wow. That's them helped out, especially in the world of the Microsoft monopoly.
This statement and others like it from the article are no different from the experience of any enthusiasts dedicated to something outside the mainstream, whether it be Amiga, Linux, Karl Popper or Tottenham Hotspur.
The phenomenom is interesting in its own right, of course, but --- as seen from the outside --- the article gives no compelling reasons to be interested in the content of this particular instance of fanaticism.
The topics seem to confuse machine consciousness with dangers of (self-replicating) nanotechnology. These are quite seperate topics in my opinion, as seperate as "is heavier than air flight possible" and "will internal combustion engines destroy the environment". Internal combustion engines are neither necessary nor sufficient for flight, and at the very speculative nature of the discussion at this stage concern with secondary effects of putting a particular implementation into practive only serves to muddy the waters.
Basically, the general problem is one of complete ignorance of the terms of the debate. The human brain is the only object in the universe which is uncontentiously involved with consciousness, and we haven't the slightest idea how it works.
Our ignorance of what consciousness is exactly is even more profound, and there are good reasons to suppose that consciousness is one of the great unanswerables, as it is the branch upon which we sit to say anything at all. To turn the saw of your analysis onto the branch that you're sitting on might only result in a fall.
There is no gaurantee at this stage that consciousness can be supported by a universal turing machine (and thus not by a computer). Hope in faster processors might be as folorn as hope in faster steam engines. Even if it could, all attempts so far at producing a viable program have been rather disappointing.
the AI community have been promising machine consciousness "in the next decade or two" for about thirty years now, and it doesn't look like we've got very far.
"THANK YOU FOR HELPING US GROW!"
How very Nietzchen. Not to mention loud.